Least-squares phase estimation with multiple parameters in phase-shifting electronic speckle pattern interferometry

We have developed an accurate and robust phase-estimation method in phase-shifting electronic speckle pattern interferometry. Unlike other methods that assume a constant phase within a fitting window, our method treats the phase variation with a gradient. A cost function that can utilize the information of pixel positions is formulated on the basis of a least-squares criterion. Powell's iteration method is applied to it to derive the phase and its gradient. An automatic consistency-checking routine and an algorithm that improves the initial guess of the iteration are developed for severe situations with large noise and steep phase variations.